I conducted factor analysis on the PCA and ran an ordinal logistic model using three factors. I've come across very interesting results.
It appears as though assets and membership are coupled as the highest loading values on factor 1 (with 47% variation explained), average balances is the highest on factor 2 (with ~25% explained) , and lending portfolios is the highest on factor 3.
Although factor 3 provides the lowest explanatory value (at ~12%) it appears to have the greatest practical significance in the model. Would I be correct in interpreting factor 3 as having the greatest practical effect on my DV regardless of the fact that it only explains 12% of the variation?
The PCA is a very helpful tool in multivariate analysis of real data. Correlations are difficult to avoid. Hopefully the PCA alone provides some insight, such as assets and membership possibly the expression of some latent variable.
I assume that by "although factor 3 provides the lowest explanatory value (at ~12%)" you mean 12% of the variance in the PCA. You are correct about factor 3. The fact that factor 1 represents the most variation in the original variables in no way implies that it is the most important predictor. Using all three factors in the model provides a strong analysis for separating and testing the effects.